[Python-checkins] cpython (2.7): Issue #25388: Fixed tokenizer hang when processing undecodable source code

[serhiy.storchaka]

https://hg.python.org/cpython/rev/8e472cc258ec
changeset:   99138:8e472cc258ec
branch:      2.7
parent:      99127:3b9fb8ebf44f
user:        Serhiy Storchaka <storchaka at gmail.com>
date:        Sat Nov 14 15:14:29 2015 +0200
summary:
  Issue #25388: Fixed tokenizer hang when processing undecodable source code
with a null byte.

files:
  Lib/test/test_compile.py |  16 ++++++++++++++++
  Misc/NEWS                |   3 +++
  Parser/tokenizer.c       |   9 ++++++---
  3 files changed, 25 insertions(+), 3 deletions(-)


diff --git a/Lib/test/test_compile.py b/Lib/test/test_compile.py
--- a/Lib/test/test_compile.py
+++ b/Lib/test/test_compile.py
@@ -3,6 +3,9 @@
 import sys
 import _ast
 from test import test_support
+from test import script_helper
+import os
+import tempfile
 import textwrap
 
 class TestSpecifics(unittest.TestCase):
@@ -555,6 +558,19 @@
         ast.body = [_ast.BoolOp()]
         self.assertRaises(TypeError, compile, ast, '<ast>', 'exec')
 
+    def test_yet_more_evil_still_undecodable(self):
+        # Issue #25388
+        src = b"#\x00\n#\xfd\n"
+        tmpd = tempfile.mkdtemp()
+        try:
+            fn = os.path.join(tmpd, "bad.py")
+            with open(fn, "wb") as fp:
+                fp.write(src)
+            rc, out, err = script_helper.assert_python_failure(fn)
+        finally:
+            test_support.rmtree(tmpd)
+        self.assertIn(b"Non-ASCII", err)
+
 
 class TestStackSize(unittest.TestCase):
     # These tests check that the computed stack size for a code object
diff --git a/Misc/NEWS b/Misc/NEWS
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -10,6 +10,9 @@
 Core and Builtins
 -----------------
 
+- Issue #25388: Fixed tokenizer hang when processing undecodable source code
+  with a null byte.
+
 - Issue #22995: Default implementation of __reduce__ and __reduce_ex__ now
   rejects builtin types with not defined __new__.
 
diff --git a/Parser/tokenizer.c b/Parser/tokenizer.c
--- a/Parser/tokenizer.c
+++ b/Parser/tokenizer.c
@@ -169,7 +169,8 @@
     tok->decoding_erred = 1;
     if (tok->fp != NULL && tok->buf != NULL) /* see PyTokenizer_Free */
         PyMem_FREE(tok->buf);
-    tok->buf = NULL;
+    tok->buf = tok->cur = tok->end = tok->inp = tok->start = NULL;
+    tok->done = E_DECODE;
     return NULL;                /* as if it were EOF */
 }
 
@@ -921,7 +922,6 @@
                 if (tok->buf != NULL)
                     PyMem_FREE(tok->buf);
                 tok->buf = newtok;
-                tok->line_start = tok->buf;
                 tok->cur = tok->buf;
                 tok->line_start = tok->buf;
                 tok->inp = strchr(tok->buf, '\0');
@@ -944,7 +944,8 @@
                 }
                 if (decoding_fgets(tok->buf, (int)(tok->end - tok->buf),
                           tok) == NULL) {
-                    tok->done = E_EOF;
+                    if (!tok->decoding_erred)
+                        tok->done = E_EOF;
                     done = 1;
                 }
                 else {
@@ -978,6 +979,8 @@
                     return EOF;
                 }
                 tok->buf = newbuf;
+                tok->cur = tok->buf + cur;
+                tok->line_start = tok->cur;
                 tok->inp = tok->buf + curvalid;
                 tok->end = tok->buf + newsize;
                 tok->start = curstart < 0 ? NULL :

-- 
Repository URL: https://hg.python.org/cpython